Expression of clinical logic with positive and negative explainability

Abstract

Methods, systems, and computer-storage media are provided for providing explainability to an end-user on why a particular patient failed to qualify for a health-related measure. High-level clinical logic explainability rules are generated on top of high-level clinical logic used to identify members of a population who qualify and who do not qualify for the health-related measure. The explainability rules are designed to identify which components in the high-level clinical logic are not satisfied by the patient's health data, and, more specifically, which of the components' sub-criteria disqualified the patient from the health-related measure. The output of the explainability rules is structured into an easy-to-understand format that is presented on a user interface associated with an end-user such as a clinician.

What is claimed is: 
     
       1. One or more non-transitory computer-storage media having computer-executable instructions embodied thereon that, when executed by a computing system comprising at least a plurality of parallel processors, facilitate a method of determining and presenting information on why members of a population do not qualify for a health-related measure, the method comprising:
 receiving one or more sets of raw patient population health data, the one or more sets of raw patient population health data received in a plurality of disparate formats; 
 automatically converting the one or more sets of raw patient population health data to one or more high-level domain-specific constructs using a system logic; 
 storing the one or more high-level domain-specific constructs in association with a distributed storage system, wherein each distributed storage is co-located with at least one of the plurality of parallel processors, wherein the distributed storage system facilitates accessing and processing of clinical information relating to a patient at substantially the same time by at least one of eliminating the need to send existing data over networks and consuming external resources; 
 accessing a machine-readable code configured to identify a population group who qualify for the health-related measure by applying a set of rules to the high-level domain-specific constructs; 
 generating permutations of the machine-readable code to produce a set of high-level clinical logic comprising at least a component corresponding to a clinical data element that is defined by one or more sub-criteria; 
 wherein the component's one or more sub-criteria comprise one or more of the following:
 a) a selection sub-criteria that specifies a portion of the grouped patient population health data that is relevant to the clinical data element corresponding to the component; 
 b) a value sub-criteria that indicates a value parameter for the clinical data element corresponding to the component that the specified portion of the grouped patient population health data must satisfy to qualify a member for the health-related measure; 
 c) a date sub-criteria that indicates a date or date range for the clinical data element corresponding to the component that the specified portion of the grouped patient population health data must satisfy to qualify a member for the health-related measure; or 
 d) a cardinality sub-criteria that indicates a minimum number of data items for the clinical data element corresponding to the component within the specified portion of the grouped patient population health data needed to qualify a member for the health-related measure; 
 executing, via the plurality of parallel processors, the set of high-level clinical logic against the one or more high-level domain-specific constructs; 
 automatically identifying, based on the executed set of high-level clinical logic, at least a first member of the population who does not qualify for the health-related measure; 
 for the at least the first member of the population, identifying, based on the executed set of high-level clinical logic, which of the component's one or more sub-criteria disqualified the first member from the health-related measure; and 
 presenting on a user interface associated with an end-user an indication that the first member did not qualify for the health-related measure, the indication including information regarding the component and the component's sub-criteria that disqualified the first member from the health-related measure. 
 
 
     
     
       2. The media of  claim 1 , further comprising:
 executing the set of high-level clinical logic against the one or more high-level domain-specific constructs; 
 automatically identifying at least a second member of the population who qualifies for the health-related measure; and 
 presenting on the user interface associated with the end-user an indication that the second member qualified for the health-related measure, the indication including information regarding the component and the component's one or more sub-criteria that qualified the second member for the health-related measure. 
 
     
     
       3. The media of  claim 1 , further comprising presenting on the user interface associated with the end-user a subset of the first member's patient population health data that corresponds to the component's sub-criteria that disqualified the first member from the health-related measure. 
     
     
       4. The media of  claim 1 , further comprising presenting on the user interface associated with the end-user a subset of the first member's patient population health data that corresponds to the component and the component's one or more sub-criteria. 
     
     
       5. The media of  claim 1 , wherein the component is part of a component expression comprising one or more additional components coupled to the component by one or more logical operators. 
     
     
       6. The media of  claim 1 , wherein the one or more sets of raw patient population health data are received from electronic medical record systems associated with one or more healthcare facilities. 
     
     
       7. A computerized method carried out by at least one server having a plurality of parallel processors for presenting information on why members of a population qualify or do not qualify for a health-related measure, the method comprising:
 receiving one or more sets of raw patient population health data, wherein the one or more sets of raw patient population health data comprises clinically-useful information in a human-readable format, the one or more sets of raw patient population health data received in a plurality of disparate media formats; 
 automatically translating the clinically-useful information of the one or more sets of raw patient population health data to one or more high-level concepts or concept groupings with simple semantics; 
 executing, using the plurality of parallel processors, a set of high-level clinical logic comprising at least a component expression comprising one or more components coupled by one or more logical operators, each component defined by one or more sub-criteria against the one or more sets of grouped patient population health data, wherein the set of high-level clinical logic is declarative such that it can be executed across the plurality of parallel processors with linear scalability; 
 wherein the execution of the set of high-level clinical logic causes at least a first member of the population who does not qualify for the health-related measure to be automatically identified and at least a second member of the population who does qualify for the health-related measure to be automatically identified; 
 for the at least the first member of the population, determining a first component of the one or more components in the component expression that disqualifies the first member from the health-related measure; 
 for the first component, identifying which of the first component's one or more sub-criteria disqualified the first member from the health-related measure; and 
 presenting on a user interface associated with an end-user:
 an indication that the first member did not qualify for the health-related measure, 
 the indication including information regarding the first component and the first component's sub-criteria that disqualified the first member from the health-related measure, and 
 an indication that the second member qualified for the health-related measure, the indication including information about the one or more components and the each component's one or more sub-criteria that qualified the second member for the health-related measure. 
 
 
     
     
       8. The computerized method of  claim 7 , further comprising:
 for the at least the first member of the population, determining a second component of the one or more components in the component expression that further disqualifies the first member from the health-related measure; 
 for the second component, identifying which of the second component's one or more sub-criteria further disqualified the first member from the health-related measure; and 
 presenting on the user interface associated with the end-user information regarding the second component and the second component's sub-criteria that further disqualified the first member from the health-related measure. 
 
     
     
       9. The computerized method of  claim 7 , wherein the end-user is a clinician caring for the first member and the second member. 
     
     
       10. The computerized method of  claim 7 , wherein the indication for the first member that did not qualify for the health-related measure further includes a subset of the first member's patient population health data that corresponds to the first component and the first component's one or more sub-criteria. 
     
     
       11. The computerized method of  claim 7 , wherein the indication for the second member that qualified for the health-related measure further includes a subset of the second member's patient population health data that corresponds to the one or more components and the each component's one or more sub-criteria. 
     
     
       12. A computer-implemented system for processing population health data to identify and explain why members of a population qualify or do not qualify for a health-related measure, the system comprising:
 a runtime engine comprising:
 a receiving component operable to:
 a) receive one or more sets of raw patient population health data associated with the members of the population, the members of the population comprising at least a first member of the population and a second member of the population, the one or more sets of raw patient population health data received in a plurality of disparate formats; 
 b) automatically group the one or more sets of raw patient population health data to one or more high-level domain-concepts or concept-groupings; 
 
 a plurality of batch processing nodes operable to:
 a) execute a set of a healthcare-domain specific declarative language against the one or more sets of patient population health data, wherein the healthcare-domain specific declarative language, utilizes high-level concepts to describe the one or more sets of patient population health data, wherein the set of healthcare-domain specific declarative language comprises at least a component expression comprising one or more components coupled by one or more logical operators, each component defined by one or more sub-criteria; 
 b) automatically identify the first member of the population, wherein the first member of the population qualifies for a health-related measure and flail the second member of the population, wherein the second member of the population does not qualify for the health-related measure; and 
 c) for the second member who did not qualify for the health-related measure, execute a set of high-level clinical logic explainability rules against the second member's grouped patient population health data to determine at least a first component of the one or more components that disqualified the second member from the health-related measure, and for the at least the first component, identify which of the first component's one or more sub-criteria disqualified the second member from the health-related measure, wherein the set of high-level clinical logic explainability rules comprise permutations of the healthcare-domain specific declarative language such that resources needed to generate the set of high-level clinical logic explainability rules are minimized; 
 
 a result interpreter component operable to at least:
 a) present on a user interface associated with an end-user, information regarding the second member that did not qualify for the health-related measure, the information comprising at least the first component and the first component's sub-criteria that disqualified the second member from the health-related measure. 
 
 
 
     
     
       13. The computer-implemented system of  claim 12 , wherein the one or more sets of raw patient population health data are received from healthcare facilities located remote to the plurality of batch processing nodes. 
     
     
       14. The computer-implemented system of  claim 12 , wherein the result interpreter component is further operable to present on the user interface associated with the end-user information regarding the first member that qualified for the health-related measure, the information comprising the one or more components and the each component's one or more sub-criteria. 
     
     
       15. The computer-implemented system of  claim 14 , wherein the information comprising the one or more components and the each component's one or more sub-criteria is presented as a textual narrative. 
     
     
       16. The computer-implemented system of  claim 15 , wherein the information comprising the one or more components and the each component's one or more sub-criteria is presented upon receiving a request from the end-user. 
     
     
       17. The computer-implemented system of  claim 12 , wherein the information comprising the at least the first component and the first component's sub-criteria that disqualified the second member from the health-related measure is presented as a textual narrative. 
     
     
       18. The computer-implemented system of  claim 17 , wherein the information comprising the at least the first component and the first component's sub-criteria that disqualified the second member from the health-related measure is presented upon receiving a request from the end-user.